The use of special observations in psychiatric practice may be employed as an alternative to more restrictive methods such as the use of seclusion and restraint. From the literature, special observations are used for a complex array of signs and symptoms (and risk behaviours) which include suicidal intent, self-injurious behaviour, hallucinatory experiences, and absconding. This paper reports on research into the use of special observations in both forensic and non-forensic psychiatric settings. A comparative approach was adopted to establish if the perceived risk factors leading to the adoption of special observations were similar in both settings. Three groups of nursing staff were requested to assess 30 patients who were placed on special observations. Before this, nurses were requested to rate the risk factors in terms of their severity on a 7-point Likert scale. The rank-ordering analysis revealed a similarity of identified risk factors and anova (one-way, unrelated) and the Jonckheere Trend Test revealed that there were significant differences between the scores in the forensic and the non-forensic settings. The statistical differences existed for risk factors relating to harm to self and others but not for psychiatric symptomatology.